Population demographics suggest that with the expected dramatic increase in age-associated dementias a public health crisis is looming. Current emphasis is on disease prevention with a focus on elderly individuals who express some cognitive impairment. We propose to identify authoritatively the risk and protective factors for cognitive decline in older persons. We have shown how to define these impaired states retrospectively, how to account for reverse transitions, how to distinguish prevalence from incidence, and how to account for competing risks by using a unique statistical (Markov) model. But sufficient longitudinal and neuropathological data is currently not available to distinguish among different types of dementia. No single Alzheimer's Disease Center (ADC) or cooperative study has an adequate sample size to reliably track transitions to dementia and differentiate Alzheimer's disease (AD) from other prevalent brain diseases that include vascular dementia (VaD) and Lewy body disease (LBD). This project will pool data from six well established longitudinal cohorts to identify risk factors for (1) preservation of intact cognition in those meeting neuropathological criteria for varying types of dementia and MCI and (2) specific forms of dementia (clinical and neuropathological). This will improve our understanding of intervening impaired states and factors that promote resistance to clinical symptoms despite the presence of neuropathology. These considerations lead to the specific aims below. Specific Aim 1: To merge databases from six large projects that follow cohorts of cognitively intact subjects to dementia, for the purpose of rigorous, statistical, biologically-informed analyses that accentuate longitudinal follow-up: BRAiNS (University of Kentucky), Nun Study (University of Minnesota), Memory and Aging Project (Washington U), Kuakini Honolulu-Asia Aging Study, Religious Orders Study (ROS, Rush Medical University), and the OHSC ADC (Oregon Health & Science University). The database would be made publicly accessible. Specific Aim 2: To identify appropriate intervening states between intact cognition and dementia based on periodic assessments of cognition and functional skills from data collected at these centers. Specific Aim 3: To study transitions and associated risk factors (e.g., genetic, medical, time in an impaired state) using novel statistical methods. Specific Aim 4: To standardize the neuropathological findings across databases (including quantitative neuropathological assessments) to enable the analysis of novel pathogenetic determinants of outcomes. This aim allows us to evaluate how actual brain pathology (e.g., microinfarcts, Lewy bodies, hippocampal sclerosis, and mixed pathologies) relates to antemortem states in the subset of participants coming to autopsy. This aim could support proposed revisions of current neuropathological and clinical research diagnostic criteria in dementia and preclinical dementia conditions.